REGISTRATION OF IMAGE SPACE WITH PHYSICAL SPACE
CT, MRI and PET scans are obtained as 3-D volumetric data
bases. There has to be registration of images with one another and with the
patient’s anatomy. Registration is divided into:
- Point Methods: these define corresponding points in
different images and physical space; determine their spatial co-ordinates
and calculate a geometric transformation between the volumes.
- Intrinsic points are anatomical landmarks intrinsic to
the patient. T his is subject to significant inaccuracy.
- Extrinsic points may be markers fixed to the skin and
rigid markers fixed to bone.
- Curve and surface methods:
Surface based registration fits sets of points extracted
from contours in one image to contours of another image or the patient’s
cranium. This has poorer registration accuracy than point based systems.
- Principal axes, correlation and interactive method: Not
used for surgery.
- Atlas method: may be combined with extrinsic point
INTRAOPERATIVE LOCALIZATION DEVICES.
Active Robotic Arm: The PUMA industrial robot was utilized
after training sessions. To define its movements and algorithms . Th bulk of
the PUMA proved restrictive in a surgical setting.
Passive arms: The ISG wand utilized electrogonio meters for
angular detectors and counterbalamers. Optical encoders have replaced
goniometers as angular sensors.
Sonic digitizer triangulation: Spark plug ultrasonic
emitters wee attached to an operating microscope in a geometric configuration.
Sound detectors for ultrasonic noise were arranged in a 3-D registered space.
The microscope focalpoint was an end affector for the assembly. This system is
sensitive to temperature and humidity.
Passive stereoscopic video: A 3-D position of an object in
space is determined from its position on 2-D video images viewed from two
different angles. This system is sensitive to line of sight considerations and
maintenance of the spatial relationship between the video camera and the
LED optical triangulation.
If 2 or more LED are attached to a surgical instrument, the
position can be traced by 3 or more cameras. These are very accurate and
flexible in usage. Line of sight considerations apply. The location of the
intraoperative localizing device is registered by external fiducial markers
and LED’s fixed to a base ring around the patient’s cranium.
Magnetic field guidance: A magnetic field is established by
a transmitter antenna near the patient’s head and its strength is measured by
a receiving antennae mounted on the surgical instrument. This system is less
accurate owing to field dislocation in the theatre as well as electromagnetic
Computer video display of Images.
A high resolution monitor with edor graphic overlays is the
standard. The on-line re-formatting of medical images is still too slow for
surgical usage and creation of segmented shaded surface and rendered images
involve time delays.
3-D localization is important for "visual fusion" in the
surgeon’s mind. As technology progresses, so will display of imaging.
Real-time Intraoperative feedback.
Digital scan information is still histomical and can become
outdated during surgical manipulation of tissues. By digitizing the video
input from intraoperative visualization, ultrasonaograpy, endoscopy, TCD and
electrophysiologic recordings may be spatially registered. Intraoperative CT +
MRI provide resolution imaging for real time updating of the surgical field.
The problems with CT and MRI intraoperatively will be dealt with later in the
APPLICATION OF NEURONAVIGATION AND PROBLEMS ENCOUNTERED.
Brain distortion is concomitant with the practice of
neurosurgery. However, the magnitude of such distortion, the influence of
tumour type and the imaging characteristics that predict shift are poorly
understood. The impact of brain shift on image guidance, the need for
intraoperative image updating and the resolution necessary for such updating
are unresolved issues.
Various techniques have been adapted to minimize this source
Positioning of the craniotomy in the horizontal plane.
Avoidance of intraoperative osmotic agents.
Attendance of sites most vulnerable to displacement early in
Placement of readiopaque beads within the tissue to enable
tracking of deformation during surgery.
Roberts et al, found a number of important observation in
their study of intraoperative brain shift:
The cortical surface moved as much as 1cm relative to
Intraoperative movement of the brain was found to be
greatest along the downward axis regardless of head position.
This was considered with the role of gravity in settling
caused by unintentional CSF drainage or soft tissue removal during surgery.
Positioning the head so that the craniotomy was superior and horizontal served
primarily to minimize lateral shift but not reduce vertical motion. The degree
of surface shift was unaffected by the size of the craniotomy.
Shift and deformation below the pial surface was not
quantified in Robert’s study. Movement of deep structures is not as great as
superficial landmarks, although quan measurement of absolute and relative
displacement of deeper structures has been limited to the observation of
Bucholz et al, using ultrasound and preoperative MRI.
Dorward et al found that during meningioma surgery, flap
positioning and tumour margin delineation can be relied upon, but the deep
tumour margin will be elevated toward the surgeon and encountered sooner than
on the neuronavigation. Shift of deep tumour margin was less in gliomas than
in meningiomas. Where a large mass was present with marked deviation of the
midline, caution should be exercised in image guided resection.
BRAIN SHIFT MEASUREMENT TECHNIQUES:
Structures examined:- skull surface at the centre of the
Cortical surface at the centre of the dural opening.
-deep tumour margin
Cortical surface adjacent to the resection margin at
A pointer tip was touched to a structure and the image
displayed on preoperative images. When shift was present, the pointertip
position appeared to lie at a distance from the chosen structure in the
images. This distance was then measured with a caliper.
Displacement occurred soon after elevation of the bone flap.
Image guided systems allow accurate placement of bone flaps and determination
of critical cortical topography with minimum exposure, but its value declines
towards the end of the case. Cranial base lesions are fixed and more rigid
than the brain and are the exception in this rule.
Overcoming the problem of brain shift and managing it more
adequately is the current forms of neuronavigation. Debulking of tumours,
aspiration of cyst and evacuation of haematomas causes a tendency of the walls
of the resultant cavity to collapse on each other, degrading the correlation
between observed and originally imaged boundaries.
is safe and easily applicable to digitizing techniques. The difficulty lies in
accurately reformatting in real time, a preoperative axial data set base on
ultrasongography, which is a medium of very different imaging characteristics
and is obtained in small sub units. Bucholz etal have created a 2 D model of
the brain, which can correct for shift on a particular plane.
Intra-operatively ultrasound is coupled for a stereotactic system. Real time
ultrasound is used to locate specific intracranial structure. Brain shift is
demonstrated by comparing preoperative images to intraoperative images.